customaize-agent:create-skill

Guide test-driven creation and validation of agent skills with YAML frontmatter and RED-GREEN-REFACTOR checklists.

Updated Apr 6, 2026
One-click install
npx skills add https://github.com/luicabref97/sushi-jungle-web --skill customaize-agent-create-skill-luicabref97
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: customaize-agent:create-skill
Source: https://github.com/luicabref97/sushi-jungle-web/tree/main/.agents/skills/customaize-agent-create-skill
Command: npx skills add https://github.com/luicabref97/sushi-jungle-web --skill customaize-agent-create-skill-luicabref97

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Provide a repeatable, test-driven framework for authoring, validating, and deploying modular Claude skills so they reliably teach procedural knowledge and resist rationalization under pressure. Solve discovery and token-efficiency issues by prescribing CSO-friendly metadata and progressive-disclosure resource loading.

Core Features & Use Cases

  • TDD-driven skill creation: RED-GREEN-REFACTOR process mapped to skill authoring, including baseline pressure tests and iterative loophole closure.
  • Discovery & metadata guidance: Strict YAML frontmatter rules, description patterns that start with "Use when...", and keyword guidance for Claude Search Optimization.
  • Testing with subagents: Pressure-scenario design, verbatim rationalization capture, and verification steps to ensure agent compliance.
  • Bundled resources guidance: When and how to include scripts, references, and assets and how to structure them for on-demand loading.
  • Deployment checklist and anti-patterns: Mandatory verification steps, red-flag lists, and examples of common mistakes to avoid.

Quick Start

Create a new skill that documents TDD-based skill authoring, includes a RED-GREEN-REFACTOR checklist, and verifies behavior with subagent pressure tests.

Frequently Asked Questions about customaize-agent:create-skill

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I use test-driven development to create agent skills?

Test-driven documentation for skill authoring uses a RED-GREEN-REFACTOR checklist to validate procedural workflows, capturing verbatim rationalizations via subagent pressure tests to ensure agent compliance and resist rationalization under pressure.

What is the best way to structure YAML frontmatter for Claude skill discovery?

Optimizing YAML frontmatter for discovery requires strict naming conventions, descriptions starting with 'Use when...', and targeted keywords for Claude Search Optimization. This metadata pattern improves token efficiency and ensures the skill surfaces for the right queries.

How do I pressure-test subagents to prevent rationalization?

Subagent pressure-testing involves designing specific scenarios, capturing verbatim rationalizations, and running mandatory verification steps. This process identifies loopholes in your skill's procedural logic and enforces strict agent compliance before deployment.

When should I bundle scripts and references with my agent skill?

Bundle scripts, references, and assets with your agent skill when on-demand loading is needed for token efficiency. Structure these bundled resources using progressive-disclosure patterns so the agent loads them only when the workflow requires specific procedural knowledge.

What are common anti-patterns and verification steps for deploying agent skills?

Deployment verification requires completing a mandatory checklist to catch anti-patterns such as missing metadata, vague descriptions, and unverified subagent behavior. Reviewing red-flag lists and common mistakes ensures the skill is bulletproof before release.